Neurocriminology: Implications for the Punishment, Prediction and Prevention of Criminal Behaviour

PERSPECTIVES

behaviour. Nevertheless, knowledge about NEUROSCIENCE AND THE LAW — SCIENCE AND SOCIETY individual genes may prove to be useful in improving our understanding of the mecha- Neurocriminology: implications for nisms and pathways that increase the risk of antisocial behaviour. Importantly, the environment plays an equally influential part; the punishment, prediction and indeed, some genetic variants confer risk of antisocial behaviour only in the presence of prevention of criminal behaviour particular environmental risk factors, such as abuse in early childhood9. Research in Andrea L. Glenn and Adrian Raine epigenetics10 has shown that the environment can influence how genes are function- Abstract | Criminal behaviour and violence are increasingly viewed as worldwide ally expressed in an individual (and even public health problems. A growing body of knowledge shows that criminal in specific brain areas); this finding under- behaviour has a neurobiological basis, and this has intensified judicial interest in mines traditional arguments of biological the potential application of neuroscience to criminal law. It also gives rise to determinism. important questions. What are the implications of such application for predicting Prenatal and perinatal influences. Early future criminal behaviour and protecting society? Can it be used to prevent health risk factors, sometimes in conjunction violence? And what are the implications for the way offenders are punished? with social risk factors, have been found to be associated with an increased probability Advances in neuroscience are increasing define antisocial behaviour and crime that a young infant will develop antisocial our understanding of how our biology influ- broadly, without distinguishing between and aggressive behaviour. During the pre- ences our behaviour — for both good and violent and non-violent offenders, this arti- natal and perinatal period, a number of fac- bad. The emerging field of neurocriminology cle largely concerns the broad propensity to tors may be important. Birth complications, seeks to apply techniques and principles from criminal behaviour. in combination with maternal rejection of neuroscience to improve our understanding the child in the first year of life, have been of crime, to predict crime and ultimately Genetics. Results from well over 100 associated with violent criminal offending to prevent crime. Such an approach would behavioural genetics studies with different at the age of 34 years in a study carried out have the potential economic and social designs — including twin studies, studies in Denmark11. This predictive finding has benefits that are associated with violence of twins reared apart and adoption studies been replicated in the United States, Canada, reduction, but it also raises neuroethical — have converged on the conclusion that Sweden and Finland with respect to violence concerns1. antisocial and aggressive behaviour have a in adulthood, and in Hawaii and Pittsburgh In this Perspective article, we discuss the considerable genetic basis. Estimates of the (USA) with respect to childhood antisocial current state of research in neurocriminol- variance that is attributable to genetics vary, behaviour1. Five other studies have shown ogy. We provide an overview of the neuro- but several meta-analyses place the level associations between birth complications and biological abnormalities that are associated at between 40–60%1. Heritable influences, externalizing-behaviour problems (such as with criminal behaviour and consider the with some exceptions, are broadly consist- aggression, delinquency and hyperactivity) in genetic and environmental factors that ent across gender and ethnicity3. Adoption children1. Fetal maldevelopment during the may contribute to these abnormalities. We studies in particular have the advantage of second trimester of pregnancy, as indicated highlight studies conducted to date, many being able to truly separate genetic from by minor physical anomalies in the child of which suggest that biological factors may environmental factors and provide a con- (such as low seated ears or a single palmar aid in the prediction of future crime and vio- verging line of evidence that there are herit- crease), has been associated with later violence. We then discuss implications of this able influences on antisocial and aggressive lent delinquency12 and violent offending in research in the legal system. behaviour (BOX 1). adulthood13. The association between fetal Recently, research has focused on iden- neural maldevelopment and childhood The current state of neurocriminology tifying which specific genes confer risk of aggression and adolescent conduct disor- There are now relatively extensive literatures antisocial behaviour. Several genetic vari- der may be even more pronounced when that document relationships between anti- ants that incrementally increase the risk of combined with effects of poor parenting14 social behaviour and biological functioning. antisocial behaviour have been identified4–7. or social adversity15. Criminal offending With some exceptions2, most studies are cor- Although approximately one-half of 185 and psychopathy have been associated with relational and cross-sectional, and largely do studies have reported positive findings, a another indicator of disruption in fetal not provide information on specific genetic meta-analysis revealed that no variant was development — namely, cavum septum or environmental factors that may mediate associated with aggression at the 5% level pellucidum16, which is the failed closure these relationships. However, an increasing of significance8. This finding underscores of the septum pellucidum, a process that number of prospective longitudinal stud- the idea that, as with other complex behav- normally takes place during gestation until ies are examining whether the presence of iours, the contribution of any single gene to approximately 6 months post-birth. Cavum specific biological factors, such as hormone antisocial and aggressive behaviour is likely septum pellucidum is thought to be an levels, neurotransmitter levels, physiological to be quite small. It is possible that a combi- early marker for disrupted development in indices or brain impairments, is predictive nation of a larger number of gene variants the limbic region of the brain17, which in of future offending. Because most studies substantially increases the risk of aggressive turn is associated with offending18.

54 | JANUARY 2014 | VOLUME 15 www.nature.com/reviews/neuro

Box 1 | Genetics and the intergenerational transmission of violence Increased testosterone levels have been repeatedly associated with increased aggres- Jeffrey Landrigan had been adopted at birth into a loving middle-class professional family. He was sive behaviour in adults. Caveats include nevertheless a particularly troublesome child from the beginning. This behaviour progressed from the fact that this relationship appears to be temper tantrums at 2 years of age, abusing alcohol at 10 years of age, being arrested for burglary at less evident in pre-pubertal individuals31, 11 years of age, abusing drugs as a teenager, to killing his first victim at the age of 20 years. After escaping from prison, he perpetrated his second killing and was sentenced to death. While he was and meta-analyses of this relationship find on death row in Arizona for this second homicide, another death-row inmate noticed an eerie a small effect size of R = 0.08 (REF. 34). Some resemblance between Landrigan and Darrel Hill, an inmate he had met on death row in Arkansas. It randomized, placebo-controlled crossover transpired that Hill was the biological father of Landrigan — a father Landrigan had never met. trials have shown that testosterone admin- Hill, like his son Landrigan, was a career criminal who also abused drugs and also killed twice. Hill’s istration increases aggressive behaviour in father — Landrigan’s grandfather — was also an institutionalized criminal who had been shot to adult males35, which is suggestive of a causal death by police. Landrigan’s great-grandfather was a notorious bootlegger. Hill had seen Landrigan connection, although other experimental only briefly as he hid two .38 pistols and the narcotic medicine Demerol under his baby son’s mattress studies using lower doses of testosterone 1 — an action that was unintentionally prophetic of Landrigan’s future drug abuse and violence. did not show an increase in aggressive As a fourth-generation criminal, Landrigan’s case documents not just the intergenerational behaviour36. Increased levels of testoster- transmission of violence but also illustrates how the adoption design separates the genetic influences of the biological parents from the environmental influences of the rearing home. Recent one at the ages of 10–12 years are predic- findings based on 43,243 adoptees and 1,258,826 non-adoptees unequivocally confirm that having tive of assaultive behaviour at the ages of a biological parent convicted of a violent crime raises the likelihood of criminal violence in the 12–14 years, norm-violating behaviour at the adoptee122. Taken together with findings from behavioural genetics studies that document age of 16 years and cannabis use at the age of heritability of aggression in children, adolescents and adults, these findings indicate that there is a 19 years37. Higher levels of testosterone at the genetic contribution to criminality. age of 16 years are associated with crime in adulthood38. Multiple neurotransmitter systems have Maternal nicotine consumption and personality disorder in the offspring27. been implicated in aggression39, and the alcohol consumption during pregnancy are Malnutrition in infancy is associated with best-replicated correlate of human aggres- also factors that may predispose individuals conduct problems in adolescence, a relation- sive behaviour is a low level of serotonin40. to violent offending in adulthood — find- ship that is partly mediated by low IQ28. Low levels of serotonin in cerebrospinal fluid ings that have been replicated across many Similarly, children with signs of malnutrition are a particular marker of people who show studies in several continents1,19. Even small at the age of 3 years have much higher rates impulsive aggressive behaviour41. An experi- amounts of alcohol during pregnancy (one of aggressive and antisocial behaviour at the mental manipulation that reduces serotonin drink per week) have been associated with ages of 8, 11 and 17 years29 over and above levels in the brain (that is, acute tryptophan increased childhood aggression in the off- any contribution from social risk factors. This depletion) reduced functioning of the orbit- spring20. There is current debate regarding relationship is also mediated by low IQ. ofrontal cortex during an inhibitory motor whether nicotine exposure predisposes Together, these findings suggest that a control task42, a region commonly implicated to offending by causing fetal hypoxia that number of early environmental factors may in antisocial behaviour43. However, aggres- results in brain impairment or whether this increase the risk of antisocial behaviour as sion has also been associated with reduced association is genetically mediated21,22. late as adulthood, probably via effects on monoamine oxidase A (MAOA) levels in Lead levels have been associated with juve- biological systems. the brain. MAOA is an enzyme that breaks nile delinquency and aggressive behaviour in down serotonin and other neurotransmit- at least six studies23. From a prospective view- Hormones and neurotransmitters. The steroid ters, and hence lower levels of MAOA would point, high lead levels in the mother during hormones cortisol and testosterone have been presumably result in higher serotonin lev- the first and second trimester of pregnancy the most intensively researched hormones in els44. This seemingly contradictory finding are associated with an increased risk of being relation to antisocial behaviour. Disruptions in demonstrates the need for studies that simul- arrested for violent crimes in adulthood24. the hypothalamus–pituitary–adrenal (HPA) taneously examine multiple biological mark- High dentine lead levels assessed at the ages of axis, the body’s stress response system that ers in order to obtain information about how 6–9 years have been associated with increased regulates the release of the hormone cortisol, neurotransmitters may interact with each violent offending at the ages of 14–21 years, are frequently observed in antisocial people. other to increase the risk of aggression. and poorer cognitive functioning mediates Associations between antisocial behaviour this relationship25. Some studies that carefully and cortisol levels vary depending on the type Psychophysiology. Psychophysiological dif- controlled for potential confounds such as of antisocial behaviour and other factors30. ferences have also been observed between poverty, maternal smoking, alcohol use and Psychological stress at various stages during antisocial groups and control groups. Meta- drug use have shown that these findings apply development may produce lasting changes in analyses and reviews conclude that low rest- to women as well as men24,25. Higher manga- HPA axis functioning and thereby predispose ing heart rate is probably the best-replicated nese levels in the mother during pregnancy an individual to antisocial behaviour31. Low biological correlate of antisocial and aggres- have also been associated with increased levels of cortisol in childhood are predic- sive behaviour in children and adolescents45,46. externalizing-behaviour problems (defined as tive of aggressive behaviour 5 years later, in Low resting heart rate may indicate a lack of aggressive, destructive and defiant behaviour) adolescence32. Similarly, a study showed that fear and a reduced likelihood of experienc- in children aged 8–9 years26. boys who were identified as having behaving negative affect in response to a criminal Poor nutrition in either the first or second ioural problems and who had low cortisol act45,46. Low heart rate in childhood and ado- trimester of pregnancy has been associ- levels showed more aggressive behaviour at a lescence has been associated with adult crime ated with a 2.5‑fold increase in antisocial follow‑up assessment 2 years later33. in all four longitudinal studies conducted to

NATURE REVIEWS | NEUROSCIENCE VOLUME 15 | JANUARY 2014 | 55

date45. Across these studies, low resting heart showed superior fear conditioning compared calculated aggression, have reduced amyg- rate was found to be as strong a predictor of with both antisocial adolescents who become dala volume72 and functioning73–75, whereas future antisocial behaviour as it is of current offenders and non-criminal controls60. individuals with a more impulsive, reactive antisocial behaviour45. In delinquents who form of aggression demonstrate exaggerated were arrested for a minor offence at the age Brain imaging and neurology. Reduced amygdala reactivity76. Reduced amygdala of 14 years, attenuated heart rate responses functioning in the frontal lobe of the brain volume in psychopathic individuals has been to a stressor were associated with both a is to date the best-replicated brain imaging localized to the basolateral, lateral, corti- shorter time to re‑offend as well as with a correlate of antisocial and violent behav- cal and central nuclei — regions that are greater number of re‑offences within a 5‑year iour. In particular, a meta-analysis of 43 involved in emotion processing, fear condi- period47. Another study showed that, after structural and functional imaging studies tioning and autonomic reactivity to affective multiple confounds had been controlled for, found that the largest reductions in struc- stimuli72. Of note, patients with damage to low heart rate at the age of 18 years predicted ture and function within the frontal lobe the amygdala have a reduced sense of dan- higher conviction frequency and higher levels of antisocial individuals were observed in ger, are less fearful77 and have deficits in the of violence up to the age of 50 years48. the orbitofrontal cortex, anterior cingulate recognition of fearful facial expressions78 (a Psychophysiological indicators of under- cortex and dorsolateral prefrontal cortex61. process involved in experiencing empathy). arousal — such as slow-frequency elec- The dorsolateral prefrontal cortex is associ- The association noted earlier59 between poor troencephalographic activity and reduced ated with self-regulatory processes, includ- classical conditioning in childhood and skin conductance activity — at the age of ing attention and cognitive flexibility, and crime in adulthood suggests, but does not 15 years are predictive of crime at the age may be linked to the antisocial features of prove, a causal relationship between amyg- of 24 years49. A recent meta-analysis50 has impulsivity and poor behavioural control61. dala functioning and antisocial behaviour. documented a reduced amplitude of the The anterior cingulate is involved in error Most brain imaging studies are essen- P300 event-related brain potential, which processing, conflict monitoring and avoid- tially correlational and cross-sectional, and is thought to reflect inefficient recruitment ance learning62–64. Individuals with damage until recently no longitudinal brain imaging of neural resources during information to this region are more disinhibited and research on antisocial populations has been processing, in adult antisocial populations. aggressive65, and demonstrate impairments conducted. Two recent studies have indicated Similarly, a reduced P300 amplitude at the in inhibitory control and emotion pro- the potential for neuroimaging to provide age of 11 years has been associated with cessing66,67. The ventral prefrontal cortex, incremental predictive power in predicting criminal offending at the age of 23 years. including the orbitofrontal cortex, has re‑offending. One study showed that reduced P300 amplitude predicted offending at the received particular attention given its role in functioning in the anterior cingulate during age of 23 years over and above measures of emotion processing, learning from reward a go–no‑go task in prisoners doubled the antisocial behaviour at the age of 11 years51. and punishment, and decision making68,69. likelihood of re‑arrest 3 years later79. A second Poor autonomic fear conditioning — the The possibility of a causal connection study of high-risk community males showed ability to learn associations between neutral between impaired orbitofrontal cortex struc- that reduced amygdala volume at the age of cues and aversive stimuli — is another well- ture and/or function on the one hand, and 26 years was associated with violent offending replicated correlate of adult criminal and crime and/or violence on the other, has been 3 years later18. As has been observed in other psychopathic adult offending52,53, conduct raised by neurological studies showing that biological longitudinal research, both studies disorder in children and adolescents54,55, and head injury in ostensibly normal individuals showed predictive utility of brain measures juvenile offending56. A review of 46 human precedes the onset of disinhibited antisocial over and above past history of antisocial brain imaging studies suggests that deficits in behaviour. For example, higher levels of behaviour and other confounds. fear conditioning may reflect abnormalities aggression were found in war veterans who Other longitudinal studies have shown in a common core fear network that consists had experienced penetrating head injuries that incurring brain damage increases the risk of the amygdala, insula and anterior cin- that were localized to the ventral prefrontal of criminal behaviour. A longitudinal study gulate57. Indeed, numerous brain imaging cortex70. Furthermore, neurological patients of 231,129 individuals from Sweden docu- studies find abnormalities in these areas in who had suffered from an accidental head mented a threefold increase in violent crime antisocial people, although this has been injury to the ventral prefrontal cortex show after traumatic brain injury (TBI) after adjust- disputed with respect to individuals with poor decision making, reduced autonomic ing for demographic confounds80. A prospec- psychopathic traits58 — a specific subgroup of reactivity to socially meaningful stimuli tive longitudinal study of 12,058 individuals criminal offenders. Poor electrodermal fear and psychopathic-like behaviour68. In a from Finland showed that TBI during child- conditioning at the age of 3 years is associated particularly striking example, a tumour in hood and adolescence was associated with a with convictions for criminal offences at the the orbitofrontal region preceded the onset 1.6‑fold increase in crime in adulthood after age of 23 years59. In addition to aiding in the of paedophilia in an individual; after resec- controlling for confounds; children suffering prediction of future offending, individual dif- tion of the tumour, the person’s behaviour from TBI before the age of 12 years started ferences in fear conditioning may also provide returned to normal71 (BOX 2). their criminal careers significantly earlier information about which antisocial individu- The amygdala is another brain region than those who suffered from TBI after the als may desist from future violence or be par- that is consistently identified as showing age of 12 years81. These studies demonstrate ticularly amenable to treatment. For example, altered activity in brain imaging studies of that information about brain structure and adolescents who were identified as being antisocial individuals. The type of deficit function, regardless of whether the origins likely to commit crimes in adulthood by vir- may vary in different subgroups of antisocial are neurodevelopmental or a result of a direct tue of being antisocial at the age of 15 years individuals. Adults and youths with psycho- physical insult later in life, may be of some but who did not go on to develop into adult pathic traits, who have blunted emotional use in identifying which individuals are at an criminal offenders at the age of 29 years responding and may engage in more cold, increased risk of criminal behaviour.

56 | JANUARY 2014 | VOLUME 15 www.nature.com/reviews/neuro

Box 2 | Ventral prefrontal dysfunction, paedophilia and legal responsibility thickness in two regions of the brain that have been implicated in antisocial behav- Cross-sectional brain imaging studies are correlational and cannot prove a causal association. iour — the orbitofrontal cortex and middle Individual case studies can, however, be suggestive of causality. frontal cortex83. Children exposed to high Michael was a 40‑year-old schoolteacher and past correctional officer. He was happily married to levels of lead early in life have been shown his wife and loved both her and Christine, his stepdaughter. He had no prior history of criminal or deviant behaviour. However, Michael began to change. He became uncharacteristically aggressive in adulthood to have reduced grey mat- with his wife and began taking pornography to school. His bedtime rituals with his pre-pubescent ter volume in the brain, particularly in the 84 stepdaughter, which had previously consisted of singing lullabies, became more sordid, and he prefrontal cortex . Males with a common eventually got into bed with her. He was found out and convicted of child molestation. polymorphism in the MAOA gene (which is Michael had to decide between a prison sentence and a treatment programme. He chose the present in about 30% of the population) have treatment programme but was expelled after propositioning female staff. The night before he was an 8% reduction in the volumes of the amyg- due to be transported to prison, he went to the emergency room complaining of a severe dala, anterior cingulate and orbitofrontal headache. There he continued to solicit sexual favours from staff. cortex85, which suggests that there is a causal An astute neurologist ordered an MRI scan after Michael wet his trousers without showing any pathway from genes to brain to antisocial apparent concern. The MRI revealed a tumour growing from the base of the orbitofrontal cortex behaviour. To delineate these types of causal (see the figure, which shows MRI scans of Michael’s brain at the time of the initial neurological evaluation, revealing a tumour mass displacing the right orbitofrontal cortex). After the tumour connections, future studies need to exam- was resected, Michael’s behaviour returned to normal, and he was reunited with his wife and ine the pathways by which genes and the stepdaughter. After several months of normal behaviour, his wife discovered child pornography on environment affect biological systems, and his computer. Michael was re‑examined, and it was discovered that the tumour had regrown. It was how these altered systems in turn predispose resected for a second time, and for at least 6 years after the resection Michael’s behaviour has individuals to antisocial behaviour. returned to normal1,71. A predisposition to criminal behaviour The case comes almost as close as one can get to a causal connection between ventral prefrontal is unlikely to be reduced to one or even two brain pathology and deviant behaviour — a pendulum moving from normality to brain dysfunction simple brain circuits but probably involves to paedophilia to neurosurgery to normality, and back again. In the face of the order in which multiple brain dysfunctions and multiple events occurred, was Michael responsible for his inappropriate sexual behaviour with his circuits that each give rise to different risk stepdaughter? factors for violence. Thus, the future use of brain imaging in the assessment of risk of criminal behaviour will require a much more sophisticated understanding of these circuits. Although brain imaging techniques have advanced rapidly in the past few decades, there are still many limitations to these methods86. However, with continued meth- odological improvements in neuroscience research, we will gain more information about how brain regions function together to predispose individuals to criminal Figure is reproduced, with permission, from REF. 71 © (2003) American Medical Association. Nature Reviews | Neuroscience behaviour. Although only a few prospective studies have been conducted, findings from research Remaining challenges for research on biologi- Important gaps in our knowledge on early risk factors suggest that information cal risk factors for violence and crime. In sum, remain. Very little is known about the about biological factors in youths may aid in recent years, evidence of the importance of neurobiology of regulatory crimes, and in the prediction of which individuals are biological factors in antisocial behaviour has one study showed increased cortical thick- more likely to engage in crime and violence accumulated and is being recognized as valu- ness and better prefrontal functioning, as later in life. Such information may also help able in our understanding of crime and vio- opposed to impairments, in white-collar to identify individuals who are particularly lence. With advances in neuroscience and the criminals82. A future challenge in neuro- amenable to rehabilitation. In a review of ten design of longitudinal investigations, studies criminology lies in parsing out the specific studies implementing variants of cognitive– are becoming methodologically stronger. genetic and environmental influences that behavioural therapy in individuals with anti- Taken together, it is becoming increasingly induce neurophysiological changes that social behaviour87, multiple neurobiological harder to argue that biological factors do result in the more proximal cognitive, affec- factors were predictive of treatment response not predispose some individuals to adult tive and behavioural risk factors for violence. and progress, including heart rate, hormone crime. This conclusion neither diminishes In other words, what neurobiological pro- levels and neuropsychological measures of nor replaces social and environmental per- cesses mediate the relationship between the risk taking, sensitivity to negative conse- spectives on crime causation24,29,80. Together, well-documented early social risk factors quences, impulsivity, cognitive flexibility and genetic and environmental factors shape and violence in adulthood? emotion processing. Although such initial the way that biological systems develop and A few studies have begun to explore how findings are provisional, these neurobiologi- function, and thus affect multiple complex genetic and environmental factors affect the cal characteristics could ultimately help to psychological processes that are important in brain. For example, researchers have found determine which offenders are best suited controlling and regulating behaviour and in that the adolescent offspring of mothers to specific rehabilitation programmes and behaving morally. who smoke during pregnancy have reduced are more likely to re‑integrate into society

NATURE REVIEWS | NEUROSCIENCE VOLUME 15 | JANUARY 2014 | 57

safely. A major challenge that remains to be he also knew that what he was doing was antisocial and criminal violence are also addressed is the identification of socially wrong. In the eyes of the law, Michael was usually dimensional in nature (for example, acceptable psychosocial or biological inter- legally responsible for his actions. the degree of prefrontal dysfunction and vention programmes that target biological Given that Michael would be considered the resting heart rate), although some may risk factors for criminal behaviour. legally responsible, it is difficult to argue that be categorical (for example, the presence someone with a less obvious neurobiologi- of TBI or genetic polymorphisms). Unlike The legal context cal ‘predisposition’ to offending than that in the United States, the judicial practice in Neurocriminology interfaces with the judi- of Michael — such as reduced functioning the Netherlands is guided by a five-point cial system at three main levels: punishment, of the amygdala during a moral decision- scale for assessing the degree of criminal prediction and prevention. To what extent making task, carrying a specific variant of responsibility, with evaluations including does the growing body of knowledge on the the MAOA gene or a significant but non- personality and neuropsychological testing93. neuroscience of crime and violence suggest obvious volume reduction in prefrontal Thus, although neuroscience has no current that we should rethink our approach in these grey matter — is not responsible for his definitional bearing on concepts of respon- three domains? Although it is unlikely that or her actions. In most criminal cases, the sibility, it is not without international prec- neurocriminology will result in any radical causal flow from biological risk to offend- edent to consider a revision to legal practice or swift shift in the operation of the criminal ing will never be known. All behaviour has in the United States, United Kingdom and justice system in the very near future, it is not a cause, and identifying the neural basis of a other countries so that responsibility may, inconceivable that some modest change may behaviour in an individual does not in itself in the future, be assessed on a continuum occur in these areas at some point, assuming establish that the individual had diminished using measures that include neurobiological that the field continues to develop and evolve, rational capacity89. Therefore, as the law cur- variables. as the past two decades have suggested. rently stands in the United States and other Although a sensible dividing line needs countries, the documentation of neurobio- to be drawn for practical reasons, in theory Punishment. Punishment is predicated on logical risk factors, no matter how early they one can conceive of a set of multiple neu- blameworthiness, and the extent to which originated, does not render that individual robiological and genetic influences that, we blame individuals is a function of the lacking responsibility. combined with social influences, diminish extent to which they can be held account- Despite this current legal stance, a chal- responsibility to varying degrees. To the able for their actions. Such accountability lenging question concerns whether the cur- extent that neuroscience provides reliable in a legal context is based on the concept of rent law pertaining to responsibility is in need methods to document these influences responsibility. of modest revision. This is ultimately a nor- objectively, and assuming that methodologies In this context, let us assume that, to some mative question over which there can be rea- become less expensive and quicker and easier extent, neurobiological abnormalities or sonable disagreement. Even without invoking to implement than hitherto, we anticipate insults relatively early in life predispose some the presence of biological risk factors to sug- that responsibility will eventually be concep- individuals to a life of crime and violence. We gest impaired rational capacity, it has been tualized more broadly than it is today. For also assume that offenders are not respon- argued that severe psychopaths should not example, although cognitive intelligence is sible for being exposed to these early risk be held responsible on the grounds that they the benchmark used by the law to document factors for violence. So are these offenders have no sense of moral rationality — they are the capacity for rationality, the relatively new responsible for their behaviour, and if so, to not sensitive to moral concerns and thus do fields of affective psychology and affective what degree? In the case of Michael, in whom not have the same moral sense as most people neuroscience are providing us with evidence a tumour in the orbitofrontal region pre- in society90. When one considers in addi- that emotion informs decision making68,94 ceded the onset of paedophilia, which disap- tion an increasing body of evidence showing — a finding that is not yet instantiated in peared after the tumour had been removed71 that neurobiological factors contribute to the law. Can individuals therefore be fully (BOX 2), the question was: was he responsible criminal psychopathy in adults as well as to responsible when the feeling for what is moral for his paedophilia? Currently, in the United behaviour in children with psychopathic- is diminished? What may be just as important States, an individual is deemed ‘responsible’ like traits91, such a revision perhaps becomes as knowing the difference between right and for their actions if two conditions are met: more compelling, particularly in a case in wrong when making moral decisions is hav- first, they have sufficient rational capacity; which an individual has several documented ing the feeling of what is right and wrong. As and second, they are not acting under coer- neurobiological and psychosocial risk factors recent studies have documented in psycho- cion. Rational capacity is typically interpreted for violence potential, as in the case of Donta paths, some individuals may have deficits in as whether the individual knew what he or Page (BOX 3). brain regions that are important for generat- she was doing and understood that his or her The judicial system acting in a practical ing these emotional responses (BOX 4). actions would have consequences. Michael’s world essentially conducts binary decision The facts that research in the field uni- (BOX 2) is a telling case because the temporal making: for example, in establishing inno- formly recognizes substantial affective ordering of events — from normality to brain cence versus guilt. Determination of dimin- impairments as a core feature of psychopa- tumour to paedophilic interest to tumour ished capacity in the United States similarly thy and that there is no longer any reason- resection to normality, and back again — is involves a categorical judgement on the able doubt that such affective impairment suggestive of causality in this particular case. presence or absence of a mental disability. influences behaviour95,96 raise the question However, in his own words, Michael admit- The fairness of this binary judgement can be of whether the legal system will eventu- ted, “…somewhere deep, deep, deep in the reasonably questioned. The widespread con- ally reformulate its current, long-standing back of my head, there was a little voice say- sensus of experts is that crime and antisocial concept of responsibility. For example, envi- ing ‘You shouldn’t do this’” (REF. 88). He knew behaviour are dimensional, and not categori- ronmental head injuries can change an oth- at the time of the act what he was doing, and cal, constructs92. Risk factors associated with erwise responsible individual into a person

58 | JANUARY 2014 | VOLUME 15 www.nature.com/reviews/neuro

Box 3 | Neuroscience in the courtroom in the future, but currently the value of geno- typing individuals to predict future violence Donta Page, a young African-American male, brutally raped and murdered Peyton Tuthill, a young is limited. white woman living in Denver (Colorado, USA) in 1999. The defendant was brought across state Perhaps surprisingly, endophenotypes lines to be scanned in the same positron emission tomography scanner with the same challenge such as prefrontal dysfunction and low task used in one prior study that had shown prefrontal dysfunction in murderers11. A comparison of the defendant’s brain scan with the average of 56 normal control individuals showed reduced heart rate, which reflect compound genetic activation in the ventrolateral, ventromedial and polar prefrontal cortex (see the figure, which and environmental influences, may cur- shows reduced functioning of the ventral prefrontal cortex in Page’s scan (left) compared with rently explain more of the variance in adult normal brains (right)). The author (A.R.) testified in the ensuing court case that such brain violence than any individual genotype and dysfunction, which potentially arose from documented severe physical abuse and head injuries in may have more traction in predicting future childhood, could predispose to poor decision making, lack of self-insight, lack of affect and poor violence. The literature reviewed above has behavioural controls, which in turn predisposes to callous, disinhibited behaviour. Inter-racial revealed several replicable early biological homicide is relatively rare and may have polarized the jury, who found the defendant guilty of correlates of later violence. Some studies first-degree murder with deliberation, punishable by death. In the death penalty hearing, a have shown that neurobiological markers three-judge panel accepted the reasoning that impaired capacity due to brain dysfunction — in can predict, over and above well-replicated conjunction with multiple additional biosocial predispositions to violence that included parental neglect, extreme poverty, sexual abuse, poor nutrition, low heart rate and lead exposure — had psychosocial risk factors, which individuals probably limited the defendant’s ability to appreciate the wrongfulness of his acts. He was spared will demonstrate antisocial or psychopathic 101,102 the death penalty and given life traits . The two recent imaging studies imprisonment. Donta Page Normal controls described above18,79, together with multiple This case highlights two studies that have identified psychophysi- competing perspectives on the ological and hormone predictors of future application of reductionist offending, provide some support for the neuroscience knowledge to the conclusions made in a Royal Society report practical, life‑or‑death issue of that neuroscience may have future value in criminal responsibility. If an predicting re-offending103. individual is burdened early in life with biological and social risk Despite the potential promise, and factors beyond their control, which, indeed likelihood, that neurobiology could in a probabilistic fashion, increase provide at least modest increases in pre- the likelihood of a criminal lifestyle, dictive power, methods used to predict are they fully responsible for their the potential of future re-offending in homicidal actions? Conversely, all about‑to‑be-released prisoners have so far behaviour has a cause that is never incorporated neurobiological markers founded in the brain. Just because a into the risk assessment equation. There are putative causal path has been three main reasons for this. First, the evolv- documented, should it be ing body of knowledge on neurocriminol- exculpatory? Would such exculpation erode our concept of ogy has not yet been accepted in the social moral responsibility? sciences and among practitioners. Second, neurobiological measures are less easy to Figure is reproduced, with permission, collect than behavioural, social and psycho- from REF. 1 © (2013) Pantheon. logical data. Third, there have been long- standing ethical concerns about collecting who, although cognitively capable of differ- or not such biologicalNature Reviews factors |are Neuroscience causes or biological data on offenders. This may entiating right from wrong, lacks the neural merely correlates of violence is irrelevant to change given that DNA is now collected on regulatory affective and behavioural control the issue of prediction — the fact that they all arrestees in the United States. Technical over their behaviour97. It has been suggested add predictive value is the currency of risk developments are also increasingly mak- that as neuroscience begins to offer a more assessment in prisoners who are about to be ing neurobiological risk assessments more detailed and specific account of the physical released. feasible and practical, and some, such as processes that can lead to irresponsible or Given that approximately 50% of the vari- the measurement of resting heart rate, are criminal behaviour, the public perception ance in aggressive and antisocial behaviour already incorporated into standard medical of responsibility may begin to change in the can be explained by genetic influences, a practice at the community level. same way that public viewpoints on addic- compelling case could in theory be made Any major advances in predicting future tion have shifted from addiction as a failure for using biological information to improve violence will be based not just on progress of personal responsibility towards addiction violence prediction. However, the fact that in neurocriminology but also on statistical as a disease98. molecular genetic studies have so far largely advances. Machine-learning techniques such failed to identify specific genes that can as random forest have already been docu- Prediction. If biological factors could account for more than 1% of the variance mented to improve the prediction of future predict future violence over and above in any complex behavioural trait gives con- charges of homicide or attempted homicide predictions based on social variables, even siderable pause for thought99,100. Molecular using traditionally available demographic opponents of a neuroscientific perspective genetic advances have, in theory, the poten- and social variables104. If neurocriminology on crime would have to agree that neurobi- tial to elucidate and identify specific genetic can identify replicable biological risk fac- ology has added value in this area. Whether factors that predispose individuals to crime tors that provide incremental knowledge

NATURE REVIEWS | NEUROSCIENCE VOLUME 15 | JANUARY 2014 | 59

Box 4 | Common neural circuits in moral decision making and antisocial behaviour Although criminal offending is heterogeneous in nature, a common — but do they have the ‘feeling’ of what is right and wrong? Moral denominator is that it is immoral. It is conceivable that the neural circuitry decision making is viewed as being influenced by affect68,94. This ‘moral underlying moral decision making is impaired in offenders. This moral feeling’, which is centred partly on the amygdala, is argued to be the neural circuit is broadly comprised of the polar and medial prefrontal engine that translates the cognitive recognition that an act is immoral cortex (PFC), ventral PFC, angular gyrus, posterior cingulate and into behavioural inhibition — a mechanism that functions less well in amygdala. These brain regions have substantial overlap with those regions affectively blunted antisocial individuals. Impairments to the that are found to be structurally or functionally impaired in offenders123 emotional component that comprises the feeling of what is moral are (see the figure, which shows a schematic diagram of brain regions that are viewed as a core feature of psychopaths and are also present in other activated only in moral decision making (green), regions that are impaired offenders. only in antisocial groups (red), and regions common to both antisocial Thus, if a criminal offender has documented disruption to this moral behaviour and moral decision making (yellow)). This overlap gives rise to neural circuitry and lacks the feeling for what is right and wrong, are they the ‘neuromoral’ hypothesis of antisocial behaviour, which states that fully accountable for their immoral behaviour? If this moral circuitry can some of the brain impairments that are observed in antisocial individuals be better delineated and quantified at the individual level in the future, disrupt moral emotion and/or decision making, thereby predisposing this affective metric could be entered as a mitigating factor in the individuals to rule-breaking, antisocial behaviour1. punishment phase of a trial, just as low IQ — a cognitive metric — is This raises an intriguing forensic question. There is little doubt that most currently used to establish lack of rational capacity and to excuse the violent psychopaths ‘know’ the difference between right and wrong defendant in the guilt phase of a trial.

Moral Antisocial Anterior cingulate Moral and antisocial

Posterior Posterior Posterior Medial PFC cingulate Medial PFC cingulate Medial PFC cingulate Ventromedial Ventromedial PFC PFC Ventromedial PFC

Insula Angular Superior Amygdala/ Superior gyrus Amygdala temporal Amygdala hippocampus temporal gyrus gyrus

Medial PFC Medial frontal Medial PFC cortex Inferior frontal Dorsolateral gyrus PFC Angular gyrus Angular gyrus Angular gyrus over and above the traditional variables Intervention and prevention. If neuro- in part be due toNature the treatment Reviews | Neuroscienceof clinical that are currently used in dangerousness criminology could provide even very modest conditions that are co-morbid with aggres- assessments, this would further aid violence insights into how future offending can be sive behaviour, such as attention-deficit prediction. Indeed, given that probation reduced, it would gain considerable traction hyperactivity disorder and depression, and parole decisions must be made every in the contexts of law and society in general, pharmacological intervention is also effec- day in offender populations, and assum- given that rehabilitation is a consideration in tive in children presenting solely with ing that neurobiological data can reliably sentencing criminal offenders. Research in aggressive symptoms. In adults with impul- enhance the accuracy of such predictions, this area is currently sparse, but some stud- sive aggression, treatment with selective it could be viewed as ethically questionable ies suggest that neurobiological research can serotonin reuptake inhibitors has been not to use such knowledge. However, such inform practice and provide guidelines for found to increase glucose metabolism in a development would raise several powerful future research. the orbitofrontal cortex106, suggesting a ethical concerns. The potential for future At the psychopharmacological level, potential method for improving function- extension of such predictions from offender it is known from over 45 randomized ing in regions that have been identified as populations to non-offender community controlled trials that that a wide range of deficient in antisocial populations. populations is one such concern given the medications — including atypical antipsy- Despite these findings, there appear egregious civil liberty violations that could chotics, mood stabilizers, stimulants and to be few, if any, systematic studies on arise from false positives — that is, non- antidepressants — are effective in reduc- the long-term efficacy of medications or dangerous individuals being predicted to be ing aggressive behaviour in children and their application to offender populations. at risk of committing crimes. adolescents105. Although such effects may Controversially, anti-androgen medications

60 | JANUARY 2014 | VOLUME 15 www.nature.com/reviews/neuro

such as medroxyprogesterone or Depo- neurocognitive functioning and prevent responsibility; fourth, considering the adop- Provera are thought to reduce recidivism offending, although these initial public tion of a dimensional concept of partial in sex offenders107, but well-controlled health prevention programmes require repli- responsibility; and fifth, discussing the randomized controlled trials are lacking. cation and extension. thorny neuroethical implications of this There is agreement that anti-androgens do Novel, innovative approaches to crime growing body of neurocriminology research reduce sexual drive, and in practice at least prevention through benign brain manipula- that include the potential for conceptual- eight states in the United States have laws on tion also have the potential to develop from izing crime as having a physical cause (for chemical castration. Although some have basic neuroscience research. One recent example, viewing crime as the result of argued that chemical castration violates the experimental transcranial direct-current psychological deficits), stigma and label- constitutional rights of the offender, oth- stimulation study showed that enhanc- ling (that is, the potentially harmful effects ers have countered that these medications ing neural excitability of the right lateral of identifying individuals based on early are effective, that offenders are capable of prefrontal cortex increases compliance to biological predispositions)98. In conclu- making an informed decision and that pre- social norms enforced by punishment117. sion, there is initial proof of concept that venting such informed choices that have Because crime is a failure to comply with neuroscience can become an important appropriate safeguards in place is ethically punishment-enforced social norms, and as future influence in society’s approach to the questionable108. brain imaging research has documented punishment, prediction and prevention of A more socially acceptable avenue of reduced lateral prefrontal functioning in criminal behaviour. biological intervention may lie in nutritional antisocial groups61, enhancing prefrontal Andrea L. Glenn is at the Center for the Prevention of supplementations such as omega‑3 fatty function could, as argued by others, have Youth Behavior Problems and the Department of acids. Several studies have documented implications for crime prevention, albeit at Psychology, University of Alabama, Tuscaloosa, Alabama 35487, USA. initial effectiveness in reducing antisocial a potential cost of reduced compliance to and aggressive behaviour in child and adult norms that are not sanctioned by punish- Adrian Raine is at the Departments of Criminology, 109–111 117 Psychiatry, and Psychology, University of populations , although null findings ment . Mindfulness training has also been Pennsylvania, Philadelphia, Pennsylvania 19104, USA. exist112. The only two randomized controlled experimentally shown to enhance both pre- 118,119 Correspondence to A.L.G. trials conducted in prison populations frontal and amygdala functioning , and e-mail: [email protected] have documented a 34–36% reduction in has been claimed to reduce aggression in serious offending in young offenders109,110. offenders120,121. We caution that this potential doi:10.1038/nrn3640 Published online 11 December 2013 Long-chain fatty acids are critical for brain for crime prevention is extremely prelimi- structure and function; they constitute 30% nary but logically follows from our review 1. Raine, A. The Anatomy of Violence: The Biological Roots of Crime (Pantheon, 2013). of the cell membrane and are known to of biological risk factors, legal implications 2. Tuvblad, C. et al. The genetic and environmental enhance neurite outgrowth and prolong cell and prevention measures. Many would agree etiology of decision-making: a longitudinal twin study. 113 J. Adolesc. 36, 245–255 (2013). life . Given the existence of structural and that once we can successfully treat offenders, 3. Ferguson, C. J. Genetic contributions to antisocial functional neural correlates of antisocial and important changes in the law and our social personality and behavior: a meta-analytic review from an evolutionary perspective. J. Soc. Psychol. 150, violent behaviour and the finding that poor perspective on crime will inevitably ensue. 160–180 (2010). nutrition is an early risk factor for antisocial 4. Cadoret, R. J. et al. Associations of the serotonin transporter promoter polymorphism with aggressivity, and aggressive behaviour, omega‑3 supple- Conclusions and future directions attention deficit, and conduct disorder in an adoptee mentation may prove to be modestly benefi- Neurocriminological research in particular, population. Compr. Psychiatry 44, 88–101 (2003). 5. DeLisi, M., Beaver, K. M., Vaughn, M. G. & cial for at least some subgroups of offenders. and neuroscience in general, are not yet Wright, J. P. All in the family: gene ×environment From a public-health perspective, appli- poised to make immediate changes in the interaction between DRD2 and criminal father is associated with five antisocial phenotypes. Crim. cations of neurobiological research on vio- prediction, prevention and punishment of Justice Behav. 36, 1187–1197 (2009). lence at the population level relatively early criminal offenders. It is also unclear how 6. DeYoung, C. G. et al. Variation in the catechol- O‑methyltransferase Val158Met polymorphism in life may help to prevent adult violence. In strong and how well replicated scientific associated with conduct disorder and ADHD one randomized controlled trial, low-income findings should be for their proper use in symptoms, among adolescent male delinquents. Psychiatr. Genet. 20, 20–24 (2010). pregnant mothers were provided with pre- legal cases, although most evidence can be 7. Gadow, K. D., Devincent, C. J., Olvet, D. M., natal and early postnatal home visitations entered as mitigating factors in the penalty Pisarevskaya, V. & Hatchwell, E. Association of DRD4 polymorphism with severity of oppositional defiant from nurses who gave advice on reducing phase of a capital punishment case. At the disorder, separation anxiety disorder and repetitive smoking and alcohol use and improving same time, notwithstanding difficulties in behaviors in children with autism spectrum disorder. Eur. J. Neurosci. 32, 1058–1065 (2010). nutrition. The study documented a 63% determining causality, there is increasing 8. Vassos, E., Collier, D. A. & Fazel, S. Systematic meta- reduction in the number of convictions convergence from different disciplinary analyses and field synopsis of genetic association studies of violence and aggression. Mol. Psychiatry among the 15‑year-old children of these perspectives that neurobiological influences http://dx.doi.org/10.1038/mp.2013.31 (2013). mothers114. One experimental environmental partly predispose an individual to offending. 9. Caspi, A. et al. Role of genotype in the cycle of violence in maltreated children. Science 297, enrichment programme that provided better It is our considered opinion that it would 851–854 (2002). nutrition, more physical exercise and cogni- be valuable for researchers and practition- 10. Tremblay, R. E. Understanding development and prevention of chronic physical aggression: towards tive stimulation to community children aged ers to focus efforts on: first, the develop- experimental epigenetics studies. Phil. Trans. R. Soc. B 3–5 years documented increased electrocor- ment of innovative and benign biological 363, 2613–2622 (2008). 11. Raine, A., Buchsbaum, M. S. & Lacasse, L. Brain tical arousal and autonomic orienting at the programmes for crime prevention; second, abnormalities in murderers indicated by positron age of 11 years115 and a 34.6% reduction in attempting to enhance the prediction of emission tomography. Biol. Psychiatry 42, 495–508 116 (1997). offending rates at the age of 23 years . In recidivism, with socially acceptable accu- 12. Arseneault, L., Tremblay, R. E., Boulerice, B., principle, targeted investment of resources racy, by including neurobiological predic- Seguin, J. R. & Saucier, J. F. Minor physical anomalies and family adversity as risk factors for violent to underserved populations at risk of future tors; third, including emotion alongside delinquency in adolescence. Am. J. Psychiatry 157, violence has the potential to enhance cognition in how we legally conceptualize 917–923 (2000).

NATURE REVIEWS | NEUROSCIENCE VOLUME 15 | JANUARY 2014 | 61

13. Mednick, S. A. & Kandel, E. S. Congenital 36. O’Connor, D. B., Archer, J., Hair, W. M. & Wu, F. C. 60. Raine, A., Venables, P. H. & Williams, M. Better determinants of violence. Bull. Am. Acad. Psychiatry Exogenous testosterone, aggression, and mood in autonomic conditioning and faster electrodermal half- Law 16, 101–109 (1988). eugonadal and hypogonadal men. Physiol. Behav. 75, recovery time at age 15 years as possible protective 14. Ryan, S. R., Schechter, J. C. & Brennan, P. A. Perinatal 557–566 (2002). factors agains crime at 29 years. Dev. Psychol. 32, factors, parenting behavior, and reactive aggression: 37. Tarter, R. E. et al. Prospective study of the association 624–630 (1996). does cortisol reactivity mediate this developmental between abandoned dwellings and testosterone level 61. Yang, Y. & Raine, A. Prefrontal structural and risk process? J. Abnorm. Child Psychol. 40, on the development of behaviors leading to cannabis functional brain imaging findings in antisocial, violent, 1211–1222 (2012). use disorder in boys. Biol. Psychiatry 65, 116–121 and psychopathic individuals: a meta-analysis. 15. Pine, D. S., Shaffer, D., Schonfeld, I. S. & Davies, M. (2009). Psychiatry Res. 174, 81–88 (2009). Minor physical anomalies: modifiers of environmental 38. van Bokhoven, I. et al. Salivary testosterone and 62. Holroyd, C. B. & Coles, M. G. H. The neural basis. of risks for psychiatric impairment? J. Am. Acad. Child aggression, delinquency, and social dominance in a human error processing: reinforcement learning, Adolesc. Psychiatry 36, 395–403 (1997). population-based longitudinal study of adolescent dopamine, and the error-related negativity. Psychol. 16. Raine, A., Lee, L., Yang, Y. & Colletti, P. males. Horm. Behav. 50, 118–125 (2006). Rev. 109, 679–709 (2002). Neurodevelopmental marker for limbic 39. Nelson, R. J. & Trainor, B. C. Neural mechanisms of 63. Kiehl, K. A., Liddle, P. F. & Hopfinger, J. B. Error maldevelopment in antisocial personality disorder and aggression. Nature Rev. Neurosci. 8, 536–546 (2007). processing and the rostral anterior cingulate: an event- psychopathy. Br. J. Psychiatry 197, 186–192 (2010). 40. Moore, T. M., Scarpa, A. & Raine, A. A meta-analysis related fMRI study. Psychophysiology 37, 17. Sarwar, M. The septum pellucidum: normal and of serotonin metabolite 5‑HIAA and antisocial 216–223 (2000). abnormal. AJNR Am. J. Neuroradiol. 10, 989–1005 behavior. Aggressive Behav. 28, 299–316 (2002). 64. Kosson, D. S. et al. The role of the amygdala and (1989). 41. Coccaro, E. F. Central serotonin and impulsive rostral anterior cingulate in encoding expected 18. Pardini, D. A., Raine, A., Erickson, K. & Loeber, R. aggression. Br. J. Psychiatry 155 (Suppl. 8), 52–62 outcomes during learning. Neuroimage 29, Lower amygdala volume in men is associated with (1989). 1161–1172 (2006). childhood aggression, early psychopathic traits, and 42. Rubia, K. et al. Tryptophan depletion reduces right 65. Devinsky, O., Morrell, M. J. & Vogt, B. A. future violence. Biol. Psychiatry http://dx.doi. inferior prefrontal activation during response Contributions of anterior cingulate cortex to org/10.1016/j.biopsych.2013.04.003 (2013). inhibition in fast, event-related fMRI. behaviour. Brain 118, 279–306 (1995). 19. Brennan, P. A., Grekin, E. R. & Mednick, S. A. Psychopharmacology 179, 791–803 (2005). 66. Danckert, J. et al. Goal-directed selective attention Maternal smoking during pregnancy and adult male 43. Blair, R. J. The amygdala and ventromedial prefrontal and response competition monitoring: evidence from criminal outcomes. Arch. Gen. Psychiatry 56, cortex in morality and psychopathy. Trends Cogn. Sci. unilateral parietal and anterior cingulate lesions. 215–219 (1999). 11, 387–392 (2007). Neuropsychology 14, 16–28 (2000). 20. Sood, B. et al. Prenatal alcohol exposure and 44. Alia-Klein, N. et al. Brain monoamine oxidase A 67. Hornak, J. et al. Changes in emotion after childhood behavior at age 6 to 7 years: I. dose– activity predicts trait aggression. J. Neurosci. 28, circumscribed surgical lesions of the orbitofrontal and response effect. Pediatrics 108, e34 (2001). 5099–5104 (2008). cingulate cortices. Brain 126, 1691–1712 (2003). 21. Maughan, B., Taylor, A., Caspi, A. & Moffitt, T. E. 45. Ortiz, J. & Raine, A. Heart rate level and antisocial 68. Damasio, A. R. Descartes’ Error: Emotion, Reason, Prenatal smoking and early childhood conduct behavior in children and adolescents: a meta-analysis. and the Human Brain (G. P. Putnam’s Sons, 1994). problems: testing genetic and environmental J. Am. Acad. Child Adolesc. Psychiatry 43, 154–162 69. Blair, R. J. The roles of the orbital frontal cortex in the explanations of the association. Arch. Gen.Psychiatry (2004). modulation of antisocial behavior. Brain Cogn. 55, 61, 836–843 (2004). 46. Armstrong, T. A., Keller, S., Franklin, T. W. & 198–208 (2004). 22. Jaffee, S. R., Strait, L. B. & Odgers, C. L. From Macmillan, S. N. Low resting heart rate and antisocial 70. Grafman, J. et al. Frontal lobe injuries, violence, and correlates to causes: can quasi-experimental studies behavior: a brief review of evidence and preliminary aggression: a report of the Vietnam Head Injury Study. and statistical innovations bring us closer to results from a new test. Crim. Justice Behav. 36, Neurology 46, 1231–1238 (1996). identifying the causes of antisocial behavior? Psychol. 1125–1140 (2009). 71. Burns, J. M. & Swerdlow, R. H. Right orbitofrontal Bull. 138, 272–295 (2012). 47. De Vries-Bouw, M. et al. The predictive value of low tumor with pedophilia symptom and constructional 23. Olympio, K. P. K., Gonçalves, C., Günther, W. M. R. & heart rate and heart rate variability during stress for apraxia sign. Arch. Neurol. 60, 437–440 (2003). Bechara, E. J. H. Neurotoxicity and aggressiveness reoffending in delinquent male adolescents. 72. Yang, Y., Raine, A., Narr, K. L., Colletti, P. & Toga, A. W. triggered by low-level lead in children: a review. Rev. Psychophysiology 48, 1597–1604 (2011). Localization of deformations within the amygdala in Panam. Salud Publica 26, 266–275 (2009). 48. Jennings, W. G., Piquero, A. R. & Farrington, D. P. individuals with psychopathy. Arch. Gen. Psychiatry 24. Wright, J. P. et al. Association of prenatal and childhood Does resting heart rate at age 18 distinguish general 66, 986–994 (2009). blood lead concentrations with criminal arrests in early and violent offending up to age 50? Findings from 73. Glenn, A. L., Raine, A. & Schug, R. A. The neural adulthood. PloS Med. 5, 732–740 (2008). the Cambridge Study in Delinquent Development. correlates of moral decision-making in psychopathy. 25. Fergusson, D. M., Boden, J. M. & Horwood, L. J. J. Crim. Justice 41, 213–219 (2013). Mol. Psychiatry 14, 5–6 (2009). Dentine lead levels in childhood and criminal behaviour 49. Raine, A., Venables, P. H. & Williams, M. Relationships 74. Jones, A. P., Laurens, K. R., Herba, C. M., Barker, G. J. in late adolescence and early adulthood. J. Epidemiol. between central and autonomic measures of arousal & Viding, E. Amygdala hypoactivity to fearful faces in Commun. Health 62, 1045–1050 (2008). at age 15 years and criminality at age 24 years. Arch. boys with conduct problems and callous–unemotional 26. Ericson, J. E. et al. Prenatal manganese levels linked Gen. Psychiatry 47, 1003–1007 (1990). traits. Am. J. Psychiatry 166, 95–102 (2009). to childhood behavioral disinhibition. Neurotoxicol. 50. Gao, Y. & Raine, A. P3 event-related potential 75. Birbaumer, N. et al. Deficient fear conditioning in Teratol. 29, 181–187 (2007). impairments in antisocial and psychopathic psychopathy: a functional magnetic resonance 27. Neugebauer, R., Hoek, H. W. & Susser, E. Prenatal individuals: a meta-analysis. Biol. Psychol. 82, imaging study. Arch. Gen. Psychiatry 62, 799–805 exposure to wartime famine and development of 199–210 (2009). (2005). antisocial personality disorder in early adulthood. 51. Gao, Y., Raine, A., Venables, P. H. & Mednick, S. A. 76. Coccaro, E. F., McCloskey, M. S., Fitzgerald, D. A. & JAMA 282, 455–462 (1999). The association between P3 amplitude at age 11 and Phan, K. L. Amygdala and orbitofrontal reactivity to 28. Galler, J. R. et al. Socioeconomic outcomes in adults criminal offending at age 23. J. Clin. Child Adolesc. social threat in individuals with impulsive aggression. malnourished in the first year of life: a 40‑year study. Psychol. 42, 120–130 (2012). Biol. Psychiatry 62, 168–178 (2007). Pediatrics 130, e1–e7 (2012). 52. Lorber, M. F. Psychophysiology of aggression, 77. Tranel, D., Gullickson, G., Koch, M. & Adolphs, R. 29. Liu, J. H., Raine, A., Venables, P. H. & Mednick, S. psychopathy, and conduct problems: a meta-analysis. Altered experience of emotion following bilateral Malnutrition at age 3 years and externalizing behavior Pychol. Bull. 130, 531–552 (2004). amygdala damage. Cogn. Neurospsychiatry 11, problems at ages 8, 11 and 17 years. Am. 53. Raine, A. The Psychopathology of Crime: Criminal 219–232 (2006). J. Psychiatry 161, 2005–2013 (2004). Behavior as a Clinical Disorder (Academic Press, 1993). 78. Adolphs, R. et al. Recognition of facial emotion in nine 30. Hawes, D. J., Brennan, J. & Dadds, M. R. Cortisol, 54. Fairchild, G., Van Goozen, S. H., Stollery, S. J. & individuals with bilateral amygdala damage. callous–unemotional traits, and pathways to antisocial Goodyer, I. M. Fear conditioning and affective Neuropsychologia 37, 1111–1117 (1999). behavior. Curr. Opin. Psychiatry 22, 357–362 modulation of the startle reflex in male adolescents 79. Aharoni, E. et al. Neuroprediction of future rearrest. (2009). with early-onset or adolescence-onset conduct Proc. Natl Acad. Sci. USA 110, 6223–6228 (2013). 31. van Goozen, S. H., Fairchild, G., Snoek, H. & disorder and healthy control subjects. Biol. 80. Fazel, S., Lichtenstein, P., Grann, M. & Långström, N. Harold, G. T. The evidence for a neurobiological model Psychiatry 63, 279–285 (2008). Risk of violent crime in individuals with epilepsy and of childhood antisocial behavior. Psychol. Bull. 133, 55. Fairchild, G., Stobbe, Y., van Goozen, S. H. M., traumatic brain injury: a 35‑year Swedish population 149–182 (2007). Calder, A. J. & Goodyer, I. M. Facial expression study. PLoS Med. 8, e1001150 (2011). 32. Shoal, G. D., Giancola, P. R. & Kilrillova, G. P. Salivary recognition, fear conditioning, and startle modulation 81. Timonen, M. et al. The association of preceding cortisol, personality, and aggressive behavior in in female subjects with conduct disorder. Biol. traumatic brain injury with mental disorders, adolescent boys: a 5‑year longitudinal study. J. Am. Psychiatry 68, 272–279 (2010). alcoholism and criminality: the Northern Finland 1966 Acad. Child Adolesc. Psychiatry 42, 1101–1107 56. Syngelaki, E. M., Fairchild, G., Moore, S. C., Birth Cohort Study. Psychiatry Res. 113, 217–226 (2003). Savage, J. C. & van Goozen, S. H. Fearlessness in (2002). 33. McBurnett, K., Lahey, B. B., Rathouz, P. J. & juvenile offenders is associated with offending rate. 82. Raine, A. et al. Increased executive functioning, Loeber, R. Low salivary cortisol and persistent Dev. Sci. 16, 84–90 (2013). attention, and cortical thickness in white-collar aggression in boys referred for disruptive behavior. 57. Sehlmeyer, C. et al. Human fear conditioning and criminals. Hum. Brain Mapp. 33, 2932–2940 (2012). Arch. Gen. Psychiatry 57, 38–43 (2000). extinction in neuroimaging: a systematic review. PLoS 83. Toro, R. et al. Prenatal exposure to maternal cigarette 34. Archer, J., Graham-Kevan, N. & Davies, M. ONE 4, e5865 (2009). smoking and the adolescent cerebral cortex. Testosterone and aggression: a reanalysis of Book, 58. Patrick, C. J., Venables, N. C. & Skeem, J. in Neuropsychopharmacology 33, 1019–1027 (2008). Starzyk, and Quinsey’s study. Aggress. Violent Behav. Psychopathy and Law: A Practitioner’s Guide (eds 84. Cecil, K. M. et al. Decreased brain volume in adults 10, 241–261 (2005). Häkkänen-Nyholm, H. & Nyholm, J.-O.) 39–77 with childhood lead exposure. PloS Med. 5, 741–750 35. Pope, H. G. Jr, Kouri, E. M. & Hudson, J. I. Effects of (Wiley, 2012). (2008). supraphysiologic doses of testosterone on mood and 59. Gao, Y., Raine, A., Venables, P. H. & Dawson, M. E. 85. Meyer-Lindenberg, A. et al. Neural mechanisms of aggression in normal men: a randomized controlled Association of poor childhood fear conditioning and genetic risk for impulsivity and violence in humans. trial. Arch. Gen. Psychiatry 57, 133–140 (2000). adult crime. Am. J. Psychiatry 167, 56–60 (2010). Proc. Natl Acad. Sci. USA 103, 6269–6274 (2006).

62 | JANUARY 2014 | VOLUME 15 www.nature.com/reviews/neuro

86. Wahlund, K. & Kristiansson, M. Aggression, 102. Raine, A., Lencz, T., Bihrle, S., LaCasse, L. & Colletti, P. 114. Olds, D. et al. Long-term effects of nurse home psychopathy and brain imaging — review and future Reduced prefrontal gray matter volume and reduced visitation on children’s criminal and antisocial recommendations. Int. J. Law Psychiatry 32, autonomic activity in antisocial personality disorder. behavior: 15‑year follow‑up of a randomized 266–271 (2009). Arch. Gen. Psychiatry 57, 119–127 (2000). controlled trial. JAMA 280, 1238–1244 (1998). 87. Cornet, L. J. M., de Kogel, C. H., Nijman, H. L. I., 103. Mackintosh, N. et al. Brain waves 4: neuroscience and 115. Raine, A. et al. Early educational and health Raine, A. & van der Laan, P. H. Neurobiological factors the law. The Royal Society [online], http://royalsociety. enrichment at age 3–5 years is associated with as predictors of cognitive–behavioral therapy outcome org/uploadedFiles/Royal_Society_Content/policy/ increased autonomic and central nervous system in individuals with antisocial behavior: a review of the projects/brain-waves/Brain-Waves-4.pdf (2011). arousal and orienting at age 11 years: evidence from literature. Int. J. Offender Ther. Comp. Criminol. http:// 104. Berk, R., Sherman, L., Barnes, G., Kurtz, E. & the Mauritius Child Health Project. Psychophysiology dx.doi.org/10.1177/0306624X13494694 (2013). Ahlman, L. Forecasting murder within a population of 38, 254–266 (2001). 88. Tiger Aspect Productions. Mindshock: sex on the probationers and parolees: a high stakes application 116. Raine, A., Mellingen, K., Liu, J., Venables, P. H. & brain. Channel Four (UK, 2006). of statistical learning. J. R. Stat. Soc. A 172, 191–211 Mednick, S. Effects of environmental enrichment at 89. Morse, S. J. Voluntary control of behavior and (2009). 3–5 years on schizotypal personality and antisocial responsibility. Am. J. Bioeth. 7, 12–13 (2007). 105. Pappadopulos, E. et al. Pharmacotherapy of behavior at ages 17 and 23 years. Am. J. Psychiatry 90. Morse, S. J. Psychopathy and criminal responsibility. aggression in children and adolescents: efficacy and 160, 1627–1635 (2003). Neuroethics 1, 205–212 (2008). effect size. J. Can. Acad. Child Adolesc. Psychiatry 15, 117. Ruff, C. C., Ugazio, G. & Fehr, E. Changing social norm 91. Viding, E. & McCrory, E. J. Genetic and neurocognitive 27–39 (2006). compliance with noninvasive brain stimulation. contributions to the development of psychopathy. Dev. 106. New, A. S. et al. Fluoxetine increases relative Science 342, 482–484 (2013). Psychopathol. 24, 969–983 (2012). metabolic rate in prefrontal cortex in impulsive 118. Desbordes, G. et al. Effects of mindful-attention and 92. Hare, R. D. & Neumann, C. S. Structural models of aggression. Psychopharmacology 176, 451–458 compassion meditation training on amygdala response psychopathy. Curr. Psychiatry Rep. 7, 57–64 (2005). (2004). to emotional stimuli in an ordinary, non-meditative 93. van der Leij, J. B., Jackson, J. L., Malsch, M. & 107. Lösel, F. & Schmucker, M. The effectiveness of state. Front. Hum. Neurosci. 6, 292 (2012). Nijboer, J. F. Residential mental health assessment treatment for sexual offenders: a comprehensive 119. Davidson, R. J. et al. Alterations in brain and immune within Dutch criminal cases: a discussion. Behav. Sci. meta-analysis. J. Exp. Criminol. 1, 117–146 (2005). function produced by mindfulness meditation. Law 19, 691–702 (2001). 108. Douglas, T., Bonte, P., Focquaert, F., Devolder, K. & Psychosom. Med. 65, 564–570 (2003). 94. Greene, J. D., Sommerville, R. B., Nystrom, L. E., Sterckx, S. Coercion, incarceration, and chemical 120. Himelstein, S. Meditation research: the state of the art Darley, J. M. & Cohen, J. An fMRI investigation of castration: an argument from autonomy. J. Bioeth. in correctional settings. Int. J. Offender Ther. Comp. emotional engagement in moral judgment. Science Inq. 10, 393–405 (2013). Criminol. 55, 646–661 (2010). 293, 2105–2108 (2001). 109. Zaalberg, A., Nijman, H., Bulten, E., Stroosma, L. & 121. Wupperman, P. et al. Mindfulness and modification 95. Neumann, C. S., Johansson, P. T. & Hare, R. D. The van der Staak, C. Effects of nutritional supplements on therapy for behavioral dysregulation: results from a Psychopathy Checklist-Revised (PCL‑R), low anxiety, aggression, rule-breaking, and psychopathology pilot study targeting alcohol use and aggression in and fearlessness: a structural equation modeling among young adult prisoners. Aggressive Behav. 36, women. J. Clin. Psychol. 68, 50–66 (2012). analysis. Personal. Disord. 4, 129–137 (2013). 117–126 (2010). 122. Hjalmarsson, R. & Lindquist, M. J. The origins of 96. Patrick, C. J., Cuthbert, B. N. & Lang, P. J. Emotion in 110. Gesch, S. M., Hampson, S. E., Eves, A. & intergenerational associations in crime: lessons from the criminal psychopath: fear image processing. Crowder, M. J. Influence of supplemental vitamins, Swedish adoption data. Labour Econ. 20, 68–81 J. Abnorm. Psychol. 103, 523–534 (1994). minerals and essential fatty acids on the antisocial (2013). 97. Sapolsky, R. M. The frontal cortex and the criminal behaviour of young adult prisoners. Br. J. Psychiatry 123. Raine, A. & Yang, Y. Neural foundations to moral justice system. Phil. Trans. R. Soc. Lond. B 359, 181, 22–28 (2002). reasoning and antisocial behavior. Soc. Cogn. Affect. 1787–1796 (2004). 111. Gustafsson, P. A. et al. EPA supplementation improves Neurosci. 1, 203–213 (2006). 98. Farah, M. J. Neuroethics: the ethical, legal, and teacher-rated behaviour and oppositional symptoms societal impact of neuroscience. Annu. Rev. Psychol. in children with ADHD. Acta Paediatr. 99, Competing interests statement 63, 571–591 (2012). 1540–1549 (2010). The authors declare no competing interests. 99. Frazer, K. A., Murray, S. S., Schork, N. J. & Topol, E. J. 112. Kirby, A., Woodward, A., Jackson, S., Wang, Y. & Human genetic variation and its contribution to complex Crawford, M. A. A double-blind, placebo-controlled traits. Nature Rev. Genet. 10, 241–251 (2009). study investigating the effects of omega‑3 100. Hariri, A. R. The neurobiology of individual differences supplementation in children aged 8–10 years from a FURTHER INFORMATION in complex behavioral traits. Annu. Rev. Neurosci. 32, mainstream school population. Res. Dev. Disabil. 31, Nature Reviews Neuroscience Series on Neuroscience and 225–247 (2009). 718–730 (2010). the law: http://www.nature.com/nrn/series/neurosciencelaw/ 101. Raine, A. et al. Corpus callosum abnormalities in 113. Calderon, F. & Kim, H.‑Y. Docosahexaenoic acid index.html psychopathic antisocial individuals. Arch. Gen. promotes neurite growth in hippocampal neurons. ALL LINKS ARE ACTIVE IN THE ONLINE PDF Psychiatry 60, 1134–1142 (2003). J. Neurochem. 90, 979–988 (2004).

NATURE REVIEWS | NEUROSCIENCE VOLUME 15 | JANUARY 2014 | 63